Using lifecyle manager to compare 400 reports output between 8.3 and 10.1.1.
- Maybe as most of you know, the html/excel format can not be used to do the comparation. This is a existing bug and if you use them, the status of comparation phase will stay "In Progress" for ever.
- So, you choice to use PDF format and you have 400 reports over 10 years in 8.3. Finally you found almost 80% reports are different when doing the comparation. So, you check what's the different. You finally found most of the differences are happened in the end and start of pages(long words(column value) are seprated across pages differently between 8.3 and 10.1.1 pdf output). As a result, the 8.3 pdf may has 600 pages while the 10.1.1 pdf has 598 pages. But the issue is user does not know if there is any data which he truely care about being different.
- So, you choice to use xml, oh...the xml file also does not make sense. One has additional xml elements while another does not between this two versions. And the elements is not user cares about.
- Finally, you thought csv file must be ideal because it only contains the data which user really cares about. And you start to believe there may be several report outputs different among 400 reports. But you found there are still many report outputs different because of 10.1.1 generates some additional column name in the header(the first line)
So, I think it is acceptable with point 4. For those different outputs, it is possible to use third party tool to check if the different is acceptable or not. And then run the reports again manually in both Cognos versions and double check the different.
Any thoughts/suggestion/experience is highly welcome.